local-serving
Guides OpenLLM local model serving, terminal chat, OpenAI-compatible localhost APIs, and readiness troubleshooting for `openllm hello`, `openllm serve`, and `openllm run`.
- License 许可
- Apache-2.0
- 最近更新
- 2026-09-03
- Stars
- ★ 351
- Forks
- 27
01 前言 Front Matter
- name
- local-serving
- license
- Apache 2.0
- metadata
- { "disco-role": "operating" }
- description
- Guides OpenLLM local model serving, terminal chat, OpenAI-compatible localhost APIs, and readiness troubleshooting for `openllm hello`, `openllm serve`, and `openllm run`.
- disable-model-invocation
- true
02 资源文件 5
- 参考文档references/cli-reference.md1.6 KB
- 参考文档references/troubleshooting.md1.8 KB
- 参考文档references/workflows.md1.8 KB
- 脚本scripts/build_serve_command.py3.1 KB
- 脚本scripts/check_local_server.py3.6 KB
03 SKILL.md 正文
Local Serving
Use this sub-skill when the task is about starting or diagnosing a local OpenLLM model server.
Typical triggers
openllm serve MODEL[:VERSION]openllm run MODEL[:VERSION]openllm hello- Chat UI at
/chat - OpenAI-compatible localhost
/v1API - Server readiness, port selection, or environment-variable forwarding
- Gated model access via
HF_TOKEN
What this route covers
openllm serveandopenllm runcommand semantics.- Default port behavior and the difference between browser chat and terminal chat.
--envand--argforwarding.ensure_bento-driven model selection and the meaning of missing or ambiguous model tags.- Readiness checks and common startup failures.
Read next
- references/cli-reference.md for command options and defaults.
- references/workflows.md for runbooks and examples.
- references/troubleshooting.md for startup, readiness, and model-loading failures.
- scripts/build_serve_command.py when you want a safe command planner instead of starting a server.
- scripts/check_local_server.py when you want to probe a running local server.
Boundaries
Do not route model repository management to this sub-skill. If the issue is a stale model cache, missing repository alias, or custom repo URL, switch to model-repositories. If the issue is BentoCloud deployment or cleanup, switch to the owning sub-skill.
04 同仓库更多 Skills
passnet-triton-opt
Make a single PassNet Triton kernel fast AND numerically faithful: performance model (when a replacement can win at all), block/grid/warp tuning, autotune policy, launch overhead, and per-op numeric recipes to pass the dtype baseline tolerances. Use when a pass matches and is correct but speedup below expected, or when correctness fails by small numeric margins.
passnet-solve
END-TO-END playbook for solving one PassNet sample: analyze the computation graphs, decide the optimization strategy (what to fuse, what to replace, what to leave alone), drive the iteration loop, and maximize the sample score. This is the ENTRY skill — invoke it first for any PassNet optimization task; it tells you when to use passnet-pattern-fusion, passnet-triton-opt and passnet-feedback.
passnet-skill
PassNet GPU kernel optimization via compiler passes. Design and implement Triton-based optimization passes, create pass files under ./pass_dir/, self-evaluate with pass_evaluator, and iterate to maximize GPU speedup.
passnet-pattern-fusion
Author PassNet pass files: write patterns that actually MATCH the FX graph, pick fusion regions (which ops to absorb into one Triton kernel), and structure multi-pass submissions with the mandatory shared-dispatch architecture. Use when creating or fixing pattern()/replacement_args()/replacement_func() files, when a pass "failed to match", or when deciding how to fuse multiple kernels/ops.
passnet-orchestrate
Round/batch planning and decision support for PassNet graph optimization. Use for multi-sample triage, worker allocation, eval-budget policy, and post-eval keep/revert decisions. For solving one concrete sample end to end, use passnet-solve as the entry skill; consult this skill only when you need broader planning or a specific decision gate.
passnet-feedback
Fast, deterministic feedback for PassNet work: pre-flight pattern/match verification WITHOUT burning a GPU evaluation, per-node bottleneck analysis of a sample's graphs, and eval-log parsing into per-variant status + estimated score + failure classification. Use BEFORE every GPU evaluation (check_pattern), at the START of a sample (analyze_graph), and AFTER every evaluation (parse_eval_log).